GABA-A Channel Subunit Expression in Human Glioma
Correlates with Tumor Histology and Clinical Outcome
Anja Smits1*, Zhe Jin2, Tamador Elsir1,3, Hugo Pedder1, Monica Niste ´r3, Irina Alafuzoff4, Anna Dimberg4,
Per-Henrik Edqvist4, Fredrik Ponte ´n4, Eleonora Aronica5,6, Bryndis Birnir2
1Department of Neuroscience, Neurology, Uppsala University, Uppsala, Sweden, 2Department of Neuroscience, Molecular Physiology and Neuroscience, Uppsala
University, Uppsala, Sweden, 3Department of Oncology-Pathology, Karolinska Institute, Karolinska University Hospital, Stockholm, Sweden, 4Department of Immunology,
Genetics and Pathology, Uppsala University, Uppsala, Sweden, 5Department of (Neuro)pathology, Academic Medical Center, Amsterdam, The Netherlands, 6Stichting
Epilepsie Instellingen Nederland, Heemstede, The Netherlands
GABA (c-aminobutyric acid) is the main inhibitory neurotransmitter in the CNS and is present in high concentrations in
presynaptic terminals of neuronal cells. More recently, GABA has been ascribed a more widespread role in the control of cell
proliferation during development where low concentrations of extrasynaptic GABA induce a tonic activation of GABA
receptors. The GABA-A receptor consists of a ligand-gated chloride channel, formed by five subunits that are selected from
19 different subunit isoforms. The functional and pharmacological properties of the GABA-A channels are dictated by their
subunit composition. Here we used qRT-PCR to compare mRNA levels of all 19 GABA-A channel subunits in samples of
human glioma (n=29) and peri-tumoral tissue (n=5). All subunits except the r1 and r3 subunit were consistently detected.
Lowest mRNA levels were found in glioblastoma compared to gliomas of lower malignancy, except for the h subunit. The
expression and cellular distribution of the a1, c1, r2 and h subunit proteins was investigated by immunohistochemistry on
tissue microarrays containing 87 gliomas grade II. We found a strong co-expression of r2 and h subunits in both
astrocytomas (r=0.86, p,0.0001) and oligodendroglial tumors (r=0.66, p,0.0001). Kaplan-Meier analysis and Cox
proportional hazards modeling to estimate the impact of GABA-A channel subunit expression on survival identified the r2
subunit (p=0.043) but not the h subunit (p=0.64) as an independent predictor of improved survival in astrocytomas,
together with established prognostic factors. Our data give support for the presence of distinct GABA-A channel subtypes in
gliomas and provide the first link between specific composition of the A-channel and patient survival.
Citation: Smits A, Jin Z, Elsir T, Pedder H, Niste ´r M, et al. (2012) GABA-A Channel Subunit Expression in Human Glioma Correlates with Tumor Histology and
Clinical Outcome. PLoS ONE 7(5): e37041. doi:10.1371/journal.pone.0037041
Editor: Maciej S. Lesniak, The University of Chicago, United States of America
Received February 1, 2012; Accepted April 12, 2012; Published May 17, 2012
Copyright: ? 2012 Smits et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work has been supported by funds from the Uppsala University (AS, BB), the Uppsala University Hospital (AS), the Selanders Foundation (AS), the
Lions Cancer Foundation at the Uppsala University Hospital (AS), the King Gustav V Jubilee Fund, Stockholm (MN, AS), Swedish Research Council (BB), the Swedish
Society for Medical Research (ZJ), and the KIKA (Stichting Kinderen Kankervrij) (EA). The funders had no role in study design, data collection and analysis, decision
to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: firstname.lastname@example.org
Gliomas are the most common form of primary brain tumor
with an overall incidence of about 4–5 per 100.000 persons per
year [1,2]. The majority of gliomas consist of glioblastomas, which
are highly proliferative and invasive tumors characterized by
remarkable biological heterogeneity and poor response to present
treatments . The prognosis for patients with diffuse low-grade
gliomas is more favorable, but these tumors transform into
malignant gliomas over time with eventually fatal outcome .
The etiology of gliomas is largely unknown, with exposure to high-
dose ionizing radiation as one of the few recognized risk factors
GABA is the main inhibitory neurotransmitter in the central
nervous system. The most common type of GABA receptors is the
GABA-A channel, consisting of a ligand-gated pentameric
chloride channel that is normally closed but can be opened by
GABA . Nineteen different GABA-A channel subunits have
been cloned and are grouped into eight separate subfamilies (a1–6,
b1–3, c1–3, d, e, h, p and r1–3). All nineteen subunits are
expressed in the brain . The GABA-A channel consists most
often of three types of subunit isoforms: two as, two bs and a third
type of subunit. All neurons contain GABA-A ion channels but the
channel subtypes change during development, varying also
between different brain regions and different types of neurons.
The distinctive functional and pharmacological of the GABA-A
channels are dictated by the composition of the subunits and can
be modulated by intracellular proteins .
GABA is present in high concentrations in the presynaptic
terminals of neuronal cells. In the postsynaptic terminal, brief
exposure to high concentration of GABA results in opening of
GABA-A channels and a subsequent increase in membrane
conductance known as phasic inhibition. It has been shown that
GABA-mediated signaling in the brain occurs also in a less
spatially and temporally restricted manner, through low concen-
trations in the extracellular space that result in a persistent or tonic
activation [5,8]. This tonic activation of GABA-A channels was
first identified in voltage-clamp recordings in hippocampal and
cerebellar neurons but is known to occur more generally in the
mammalian brain . Extrasynaptic pharmacologically active
PLoS ONE | www.plosone.org1 May 2012 | Volume 7 | Issue 5 | e37041
GABA-A channels in the mature brain have been found in non-
neuronal cells such as astroglial cells [10,11], and also in T
lymphocytes , suggesting an immunoregulatory role for
extrasynaptic GABA .
There is growing evidence also for a widespread role of GABA
in the growth regulation of many cell types, including neuronal
stem cells and perhaps tumor stem cells . In the postnatal
subventricular zone along the lateral ventricle where adult
neurogenesis occurs, GABAergic signaling is involved in the
proliferative control of neuroblasts . In primary brain tumors,
GABA-A channel subunits have been detected in the neuronal
component of gangliogliomas , but also in gliomas where the
response to GABA correlated with the malignancy grade of the
tumor [15,16]. GABA-evoked current response was restricted to
low-grade gliomas and not recorded in glioblastoma, suggesting
that the disappearance of GABA-A channels parallels the
unlimited growth of malignant gliomas . When co-cultured
with neurons, an interaction between neurons and glioma cells was
triggered resulting in functional expression of GABA-A channels
by tumor cells within 24 hours .
These reports are all in favor of a role for extrasynaptic GABA
in the growth control of gliomas, but so far this has not been
substantiated by studies on clinical patient material. Also, a
systematical analysis of the distribution of the different GABA-A
channel subunit isoforms in human gliomas has not been
performed. In the present work we compared mRNA levels of
all 19 GABA-A channel subunits in gliomas of different
malignancy grade and histological subtype. We then studied the
distribution of the a1, c1, r2 and h subunit proteins in vivo in a
clinical cohort of diffuse low-grade gliomas and correlated the
presence of GABA-A channel subunit proteins with patient
Quantification of GABA-A channel subunit mRNAs levels
The histological classification of the tumors samples used for
quantitative real-time PCR (qRT-PCR) is shown in Table 1. Mean
expression levels of GABA-A channel subunit mRNAs were
compared group-wise between gliomas grade II (n=12), gliomas
grade III (n=10) and glioblastomas (n=7) according to the World
Health Organization (WHO) classification of brain tumors .
Seventeen GABA-A channel subunit mRNAs were consistently
detected, whereas the r1 and r3 subunits were not expressed. The
relative mRNA levels of the 17 detectable GABA-A subunits in
gliomas of different malignancy grade are shown in Figure 1.
Statistical analysis of quantitative differences between gliomas
grade II, gliomas grade III and glioblastomas showed significantly
higher mRNA levels of a1, a6, c1 and c2 GABA-A channel
subunits in gliomas grade II compared to glioblastomas, and of a3,
a6, b3, c1, c2 and p subunits in gliomas grade III compared to
glioblastomas (p,0.05). In contrast, mRNA levels of the h subunit
were 5–10 fold higher in glioblastomas than in gliomas with lower
malignancy grade (p,0.05) (Figure 2).
For 15 of the 17 detectable GABA-A subunits, no statistically
significant differences in mRNA levels were found between
astrocytomas and oligodendrogliomas grade II (Figure S1) or
between astrocytomas and oligodendrogliomas grade III (Figure
S2). The two exceptions were the b1 subunit, for which expression
was lower in astrocytomas grade II than in oligodendrogliomas
grade II (Figure S1), and the a3 subunit, showing lower expression
in astrocytomas grade III than in oligodendrogliomas grade III
(Figure S2) (Mann-Whitney Rank Sum Test).
In peri-tumoral brain tissue, mRNAs for the 17 GABA-A
channel subunits were consistently detected (Figure S3). For most
subunits belonging to the a, b and c subfamilies, as well as for the
d subunit, mRNA levels were 3–5 fold higher in peri-tumoral
brain tissue compared to tumor samples. Other subunits were
expressed at approximately similar levels as those found in tumors
Distribution of GABA-A channel subunit proteins in
Table 2 summarizes patient characteristics and tumor charac-
teristics of the 91 gliomas that were used for immunohistochemical
analysis. The results of the immunostainings are shown in Table 3.
Tumor cells with immunoreactivity for a1, c1, r2 and h GABA-A
channel subunits were identified in all histological tumor types.
The subcellular distribution of the subunit proteins was found to
be cytoplasmic and membranous. Of the total number of 91
samples, 17 showed immunoreactivity for the a1 subunit (19%), 24
for the c1 subunit (26%), 29 for the r2 subunit (32%), and 35 for
the h subunit (38%) (Figure 3). The mean fraction of immuno-
reactive tumor cells denoted as + was between 4–8% for the a1,
c1, r2 and h subunits (Table 3). The mean fraction of
immunoreactive tumor cells in astrocytomas denoted as ++ was
15% for the a1 subunit (n=1), 20% for the c1 subunit (n=2),
55% for the r2 subunit (n=5) and 40% for the h subunit (n=9).
In oligodendrogliomas, the mean fraction of immunoreactive
tumor cells denoted as ++ was 30% for the r2 subunit (n=1) and
50% for the h subunit (n=2), while in oligoastrocytomas these
percentages were 50% for the c1 subunit (n=3), 35% for the r2
subunit (n=2), and 40% for the h subunit (n=3).
Three astrocytomas grade II showed immunoreactivity for all
four subunits, while none of the oligodendrogliomas and
oligoastrocytomas expressed all four subunits. Interestingly, more
than 80% of the gemistocytic astrocytomas expressed the r2 and h
subunits, but lacked immunoreactivity for the a1 and c1 subunit
proteins (Figure 4). A similar pattern of predominantly r2 and h
subunit expression was found in minigemistocytic oligoastrocyto-
mas (Table 3).
As a first step, we dichotomized the immunohistochemistry
results of the four GABA-A channel subunits into two categories;
no immunoreactivity (=0) versus positive immunoreactivity (+ or
++). These data are illustrated in Figure 5, where each number in
the figure represents one tumor sample consisting of astrocytomas
(n=42), oligodendrogliomas (n=27) or oligoastrocytomas (n=18)
grade II, with either no (no bar) or positive (colored bar)
immunoreactivity for the four subunits. As demonstrated, there
was a frequent co-expression of r2 and h subunits in astrocytomas
grade II (r=0.86, p,0.0001) and in oligodendrodendrogliomas/
oligoastrocytomas grade II (r=0.66, p,0.0001). No statistically
significant correlation between the a1 and c1 subunit or between
any of the other four subunits was found.
The mean postoperative survival for patients with gliomas grade
II (n=87) was 7,9 y. As expected, mean survival was longer
amongst patients with oligodendrocytomas compared to oligoas-
trocytomas and astrocytomas (11.6 y, 6.9 y and 5.9 y respectively).
We hypothesized that r2 expression and h subunit expression in
grade II gliomas were prognostic factors for survival and tested
these two parameter as dichotomized variables in the survival
GABA-A Channel in Human Glioma
PLoS ONE | www.plosone.org2 May 2012 | Volume 7 | Issue 5 | e37041
analysis for astrocytomas grade II (n=42) and for the combined
group of oligodendrogliomas and oligoastrocytomas (n=45).
Both r2 subunit expression and h subunit expression were
associated with statistically significant longer median survival in
the histological subgroup of astrocytomas (Table 4). Of the
established clinical and radiological prognostic factors for grade
gliomas grade II (age, epilepsy, KPS, tumor size, tumor location,
contrast enhancement) , the parameters seizures at onset,
KPS$90 and tumor diameter #6 cm were identified as prognos-
tic factors in the univariate analysis (Table 4). Young age (#40
years), non-enhancing tumor and non-central tumor localization
were not associated with favorable prognosis. The presence of
mutated IDH1 R132H protein showed a trend towards longer
survival in the univariate analysis and was entered as a variable in
the multivariate model . Since all patients but one were treated
by radiotherapy, this parameter was not tested. Likewise, the LOH
1p/19q status was positive for only two patients and could not be
entered into the model. As shown in Table 4, expression of r2
subunits contributed to significantly longer survival in the
multivariate model, together with seizures at onset, KPS$90,
and mutated IDH1 protein. Expression of h subunits was not
identified as an independent prognostic factor in the multivariate
model. Small tumor size was a significantly favorable prognostic
factor entered as a single variable (p=0.019), but lost its
significance in the multivariate model.
In the group of oligodendrogliomas and oligoastrocytomas
(n=45), the parameters oligoastrocytoma histology (versus oligo-
dendroglioma) and tumors crossing midline structures (versus non-
central tumor location) were identified as unfavorable prognostic
factors in both the univariate and multivariate model (Table 4).
The parameters expression of r2 subunit and expression of h
subunit were not associated with significantly longer survival in
The aims of the present study were 1) to explore the distribution
of GABA-A channel subunits in human gliomas of various
histological subtypes, and 2) to search for correlations between
GABA-A channel subunits and patient survival. We showed that
17 of the 19 different GABA-A channel subunit mRNAs were
Figure 1. Summary of the qRT-PCR results showing mRNA levels of 17 different GABA-A subunits in gliomas grade II (n=12),
gliomas grade III (n=10) and glioblastomas (n=7). The normalized mRNA expression of each target gene relative to a reference gene TATA-
binding protein (TBP) was calculated using the 22DCtmethod.
Figure 2. Detailed presentation of the RT-PCR results for each of the 17 subunits, showing quantitative mRNA levels between
gliomas grade II, gliomas grade III and glioblastomas. Statistically significant differences in mRNA levels are marked (*). The normalized mRNA
expression of each target gene relative to a reference gene TATA-binding protein (TBP) was calculated using the 22DCtmethod.
GABA-A Channel in Human Glioma
PLoS ONE | www.plosone.org3 May 2012 | Volume 7 | Issue 5 | e37041
present, only r1 and r3 subunits could not be detected. While no
major differences in subunit mRNA levels were found between
astrocytomas and oligodendrogliomas, there was a general down-
regulation of GABA-A channel subunits in glioblastoma. Thus,
highest mRNA levels were recorded in gliomas grade II and grade
III, except for the h subunit that showed 5–10 fold higher levels in
Our findings are in agreement with previous reports demon-
strating functional bindings sites for the GABA-A channel receptor
in low-grade and anaplastic gliomas but not in glioblastomas
[16,17]. The lack of responsiveness to GABA in glioblastomas has
been explained by either the loss of GABA-A channels or by
defective channels . Thus, tumor cells equipped with
functional GABA-A channels may be able to maintain low or
intermediate proliferative activity upon physiological levels of
GABA in the extracellular fluid. GABA triggered a depolarization
in the majority of glioma cells and a concomitant increase of
intracellular Ca2+. Since the GABA-A channel itself is not
permeable for Ca2+ions, the Ca2+influx has been associated with
an activation of voltage-gated Ca2+channels in the tumor cell
membrane. The central role of intracellular Ca2+as a downstream
effector of GABA receptor signaling in cell proliferation has been
The other main finding of this study is the strong co-expression
of r2 and h subunit proteins in gliomas grade II and the
association between r2 subunit expression and longer patient
survival in diffuse astrocytomas. In spite of its strong co-expression
with the r2 subunit, the h subunit did not emerge as an
independent predictor of survival in the multivariate model. These
findings are in agreement with the qRT-PCR results, showing an
opposite trend for the h subunit with up-regulation in glioblasto-
mas. The prognostic impact of r2 subunits in diffuse astrocytomas
is intriguing but is not necessarily unique for the GABA A-channel
subunit families. It is possible that we failed to identify associations
for other subunits in oligodendrogliomas and oligoastrocytomas,
which were combined into one group in the present study due to
the limited number of tumor samples. Neither can we exclude that
there has been a selection bias influencing our results, since many
tumor biopsies from the original clinical cohort were too small to
be included in the TMA. Thus, our study provides a first link
between GABA-A channel composition and survival in gliomas
and supports a presumed functional role of GABA in gliomas, but
the findings need to be confirmed in larger studies.
The GABA-A channel subunits exist as a family of subtypes
with distinct temporal and spatial patterns of expression and
distinct properties that presumably underlie a precise role for each
subtype [21,9]. The differential sensitivity to GABA is dependent
on the specific subunit composition of the A-channel, and may
further be influenced by intracellular proteins interacting with the
channel complex and by post-translational modifications .
Whilst the a1 subunit has frequently been the focus of studies, due
to its abundance in the brain and its reliability of measurement,
considerably less is known on the other subunits expressed at a
lower level such as h and the r2 subunits. In the brain, functional
GABA-A channel receptors containing the h subunit are formed
together with a, b and c subunits . Synaptic channels are
activated by high GABA concentrations, whereas extrasynaptic
channels can be activated by extremely low concentrations of
GABA. Tumors such as gliomas are probably not exposed to high
GABA concentrations and one would expect mostly tonic currents
to be active in these tumors. It seems that tonic currents more or
less can contain any subunit in their channel complex, whereas
synaptic channels are more restricted to certain types of subunits.
As such, the c2 subunit is a major component of synaptic channels,
while the a4 a5 a6 and d subunits are part of extrasynaptic GABA
A-channels and thus involved in tonic currents [7,23]. The a1, a2
and a3 subunits can be located both at and outside of synapses and
the e and h subunits are probably mostly extrasynaptic [7,23].
In the present study, r2 subunit expression was found in a
relatively large number of glioma samples while no detectable
mRNA levels were found by qRT-PCR for the r1 and r3
subunits. The r1 subunit is predominantly expressed in the retina
and visual pathways, while r2 expression has also been found in
hippocampus and amygdala [24,25]. Rho subunits form functional
homomeric GABA-A-r receptors (previously known as GABAC
channels) [7,26], but can potentially also participate in hetero-
meric GABA-A channel formation [7,27]. The GABA-A-r
receptor has a number of distinctive and unique functional
properties, such as a long mean opening time of the channel and
slow desensitization, suggesting a more widespread function than
previously thought . In addition, the pharmacological
properties of the GABA-A-r receptor, including the lack of
response to benzodiazepines and barbiturates, set apart this class
Table 1. Histopathological diagnoses of the samples
included in the qRT-PCR (n=33).
DiagnosisNumber of samples
Gliomas grade II 12
Gliomas grade III 10
Figure 3. Photomicrographs of an oligodendroglioma grade II
showing positive immunoreactivity for: A) GABA-A channel a1
subunit, and B) c1 subunit. The histological markers used for
identification of specific cell types, CD34, IDH1, Ki67 and neuronal nuclear
antigen (NeuN) illustrate respectively, C) haematopoetic cells in the tumor,
of proliferating tumor cells, F) an entrapped neuron in the tumor. (The scale
bar shown in A represents 50 mm in A, B, F; 80 mm in C, D, E).
GABA-A Channel in Human Glioma
PLoS ONE | www.plosone.org4 May 2012 | Volume 7 | Issue 5 | e37041
We found a strong positive correlation between the expression
of h and r2 subunits. Expression of the h and r2 subunits in brain
tissue is relatively uncommon, and there is no evidence to suggest
that they form a functional channel together [29,30]. However,
normal transcriptional regulation in tumor cells is disrupted and
may not follow the normal ‘‘rules’’ for GABA-A channel subunit
expression. It is therefore possible that they could have been
expressed without participating in forming functional channels.
Interestingly, the strongest correlation between the expression of
r2 and h subunits was found in gemistocytic astrocytomas. The
presence of gemistocytes in a tumor has been correlated with
worse prognosis [31,32], although the proportion of gemistocytes
in the tumor does not seem to have an effect . The reasons for
development of gemistocytes in tumors are still poorly understood,
but it has been suggested that they may develop due to losing the
competition for substrates with adjacent cells [31,32].
It is not known whether tumor cells are the source of GABA.
Only neurons contain glutamic acid decarboxylase (GAD), the
enzyme that catalyzes the decarboxylation of glutamate to GABA,
and are able to produce GABA. However, astrocytes can take up
GABA and there is a minor pathway not involving GAD that
allows synthesis of GABA . Both neurons and astrocytes can
release GABA, but only neurons can release high concentrations
of GABA within a short time-span from synaptic vesicles at the
presynaptic terminal .
The present study shows a down-regulation of subunit mRNA
levels in glioblastomas, except for the h subunit, and the presence
of distinct GABA-A channel subunit proteins in gliomas grade II.
The correlation between r2 subunits and favorable survival in
diffuse astrocytomas suggests that specific subunit compositions
affect clinical outcome in glioma. Taken the distinct pharmaco-
logical properties of the different GABA-A channel subtypes, this
may open up for new therapeutic strategies for glioma. Our study
also highlights the complexity of GABA signaling in gliomas and
stresses the need for functional studies in this field.
Materials and Methods
Informed consent was obtained for the use of human brain
tissue and for access to medical records for research purposes, and
all material was obtained in a manner compliant with the
Declaration of Helsinki. The studies involving human tissue
samples were approved by the Ethics Committee of Uppsala
University (Application Dnr Ups 02-330) and the Ethics Com-
mittee of Karolinska Institutet (Application Dnr Ki 02-254).
Written informed consent was obtained prior to sample collection.
Table 2. Clinical and tumor characteristics of gliomas included in the TMA (n=91).
Parameter Number of patients (%)
Male 56 (62)
Mean age (at onset) 40.3 y
Others 15 (16)
Biopsy only 15 (16)
Subtotal resection 50 (55)
Gross total resection26 (29)
Frontal 50 (55)
Fronto-temporal/fronto-parietal 11 (12)
Temporal 15 (17)
Temporo-parietal/parieto-occipital 4 (4)
Central 2 (2)
HistologyTotal LOH 1p/19qMutated IDH1 protein
Astrocytoma grade II 422 32
Oligoastrocytoma grade II183 17
Oligodendroglioma grade II 272024
Astrocytoma grade III100
Oligoastrocytoma grade III313
GABA-A Channel in Human Glioma
PLoS ONE | www.plosone.org5 May 2012 | Volume 7 | Issue 5 | e37041
Tissue Samples for quantitative real-time quantitative
Human tissue samples for quantitative real-time quantitative
PCR (qRT-PCR) were obtained from the files of the department
of neuropathology of the Academic Medical Center, University of
Amsterdam. Histological classification of tumors according to the
WHO is shown in Table 1. Twenty-nine gliomas (12 gliomas
grade II, 10 gliomas grade III, 7 primary glioblastomas), all with
temporal lobe location, were included. Five samples of peri-
tumoral brain derived from patients with primary glioblastomas in
the temporal lobe were included. These samples consisted of
tumor-free cortex on the basis of absence of immunoreactivity for
Ki67 (1:500; Rabbit polyclonal, DAKO), IDH1 R132H (1:100;
Clone H09, Dianova) and p53 (1:2000; Clone DO-7+BP53-12,
Neomarkers). Samples were quickly removed at surgery and
immediately divided into two parts; one part was fixed in 4%
paraformaldehyde for 24 hours, paraffin embedded and used for
histopathological diagnosis, the remaining part was snap frozen in
liquid nitrogen and maintained at 280uC until used for RNA
Quantitative real-time PCR
Total RNA was isolated with TRIzol reagent according to the
manufacture’s instruction and further transcribed to cDNA using
oligodT primers (Table S1). qRT-PCR was performed in a 10 ml
reaction mixture containing 4 ml cDNA (0.5 ng/ml), 16 PCR
Table 3. Results of immunostaining for GABA-A channel subunits on gliomas (n=91).
Subunit/HistologyNr of samples (n)
Nr of positive tumor cells*
Astrocytoma grade II30 231922 210
Gemistocytic astrocytoma12 1112220
Oligoastrocytoma grade II 15138 14120
Oligodendroglioma grade II 27 232122 180
Astrocytoma grade III101000
Oligoastrocytoma grade III323220
*0=No immunoreactive tumor cells; +=,10% immunoreactive tumor cells; ++=$10% immunoreactive tumor cells.
Figure 4. Photomicrographs of a gemistocytic astrocytoma grade II showing no immunoreactivity for: A) and A9) GABA-A channel
a1 subunits, B) and B9) c1 subunits, but positive immunoreactivity for: C) and C9) r2 subunits and D) and D9) h subunits in a
proportion of the tumor cells. (The scale bar shown in A represents respectively 200 mm in A, B, C, D; 30 mm in A9, B9, C9, D9).
GABA-A Channel in Human Glioma
PLoS ONE | www.plosone.org6 May 2012 | Volume 7 | Issue 5 | e37041
reaction buffer, 3 mM MgCl2, 0.3 mM dNTP, 16 ROX
reference dye, 0.8 U JumpStart Taq DNA polymerase (Sigma-
Aldrich), 56 SYBR Green I (Invitrogen) and 0.4 mM each of
forward and revere primers. The gene-specific primer pairs were
designed using Primer Express Software version. 3.0 (Applied
Biosystems), synthesized by Invitrogen and further validated using
BioBank cDNA from human brain (PrimerDesign). The specific
primer sequences for the 19 different GABA-A subunits are shown
in the supplementary material (Table S1). Amplification was
performed in 96-well optical plates using the ABI PRISM 7900HT
Sequence Detection System (Applied Biosystems) with an initial
denaturation of 5 min at 95uC, followed by 45 cycles of 95uC for
15 s, 60uC for 30 s and 72uC for 30 s. A melting curve was
determined at the end of cycling to ensure the amplification of a
single PCR product. Each reaction was run in duplicate. Cycle
threshold values (Ct) were determined with the SDS 2.3 software
supplied with the instrument. The expression of each target gene
relative to a reference gene TATA-binding protein (TBP) was
calculated using the 22DCtmethod as previously described .
Tissue samples for immunohistochemistry
The distribution of GABA-A channel subunit proteins was
explored by immunohistochemistry in a tissue microarray (TMA)
of glioma samples. The TMA was prepared from paraffin-
embedded tissue blocks as previously described . Samples were
included from patients aged $16 years old operated for low-grade
gliomas between 1984–2001 at the Uppsala University Hospital,
as part of a previously described clinical cohort of diffuse low-
grade gliomas . Paraffin-embedded tissue blocks were identi-
fied and re-used for preparation of the TMA. All tumors included
in the TMA were re-evaluated for the purpose of the study by one
of the neuropathologists (IA). Of the original cohort of 116 grade
II gliomas, 3 tumors were re-classified as non-gliomas, 18 blocks
contained too little tumor for TMA preparation (mainly biopsies),
and in 4 blocks areas with infiltrative tumor edge were left without
representative areas of tumor bulk. Thus, 91 glioma samples
remained and were included in the TMA presented in this study.
Clinical and histopathological tumor characteristics are presented
in Table 3.
Figure 5. Illustration of the immunohistochemistry results in astrocytomas grade II (n=42), and in oligodendrogliomas (n=27) plus
oligoastrocytomas grade II (n=18). Positive immunoreactivity for each of the four different GABA-A channel subunits is visualized by a colored
bar. Each number represents one tumor sample.
GABA-A Channel in Human Glioma
PLoS ONE | www.plosone.org7 May 2012 | Volume 7 | Issue 5 | e37041
Clinical data collection
A retrospective chart review was performed for all tumor cases
as previously described . The following clinical data were
collected: time-point of first symptoms, patient age at disease
onset, date of operation, date of death, tumor size, tumor location
(specific lobe, and (sub)cortical versus central location), date and
extent of diagnostic surgery (biopsy, subtotal or gross total
resection). The extent of tumor resection was based on postop-
erative CT scans or on the surgeon’s operative notes. Survival was
defined as the time point between operation and date of death or
end of the study (20 September 2009). Data concerning time of
death and the cause of death were collected from central health
authorities (the National Cause of Death Register Data).
Table 4. Prognostic factors for survival in astrocytomas grade II (n=42) and oligodendrogliomas/oligoastrocytomas (n=45) using
Log-Rank model (univariate) respectively Cox’s proportional hazard model (multivariate).
ParameterNr median survival95% CI p-valueRR 95% CIp-value
ASTROCYTOMAS42 4.4 3.7–6.5
Seizure at onset:
no7 2.00.5–2.3 0.000312.73.5–45.60.0002
yes 354.8 4.1–7.5
Age at onset:
,40 y 154.3 1.1–8.2 0.73
#40 y27 4.73.4–7.2
$90267.5 3.9–8.6 0.0540.3 0.2–0.80.009
$6 cm 14 3.41.5–5.70.01931.1 0.5–2.80.89
,6 cm28 5.34.1–8.5
yes 325.6 3.7–7.40.0940.29 0.1–0.80.020
yes 196.5 4.0–8.60.023 1.50.2–5.50.64
no 234.0 2.7–4.9
yes8 4.52.2–8.40.002 4.31.7–10.10.003
no 379.2 6.4–11.1
no36 8.4 4.7–9.6
GABA-A Channel in Human Glioma
PLoS ONE | www.plosone.org8 May 2012 | Volume 7 | Issue 5 | e37041
Tissue microarray preparation
TMAs were prepared by the Swedish Proteome Resource
Centre (HPR) facilities at the department of pathology, the
Rudbeck laboratory, Uppsala University Hospital . In brief,
donor blocks were sectioned, hematoxylin-eosin stained and
immunostained with immunohistochemical markers to select
appropriate areas for the TMA construct. Doublets of represen-
tative areas were included for each tumor. Using an automated
tissue arrayer (Beecher Instruments, Silver Spring, MD), two 1.0-
mm diameter punches were taken from each donor block and
transferred to the recipient TMA block. In total, two TMA blocks
Prior to immunostaining, sections were deparaffinized, hydrated
in graded alcohol, and microwave treated for 10 min at 750 W
and 15 min at 350 W. Heat-induced epitope retrieval was
performed by heating the TMA slides immersed in retrieval
buffer pH 6 (Lab Vision, Freemont, CA) for 4 min at 125uC in a
pressure boiler (Decloaking chamber, Biocare medical). Immuno-
staining was performed by avidin-biotin peroxidase staining
technique (Vector elite), using 3,3-diaminobenzidine as a sub-
strate. The presence of heterogeneous cell populations in the TMA
was documented using the following immunohistochemical
markers: GFAP (1:500; polyclonal rabbit, DAKO, Denmark),
vimentin (1:80; Mouse clone V9, Sigma), neuronal nuclear protein
(1:1000; Mouse clone MAB377, Chemicon, Temecula, CA),
microtubule-associated protein (1:100; Mouse clone HM2, Sigma),
CD34 (1:100; Mouse monoclonal, DAKO), Ki67 (1:500; Rabbit
polyclonal, DAKO), and mutated isocitrate dehydrogenase 1
(IDH1) R132H protein (1:100; monoclonal mouse antibody,
Fluorescent in situ hybridization
All samples included in the TMA were studied by fluorescent in
situ hybridization analysis (FISH) to identify losses of chromosomal
arms 1p and 19q (LOH 1p/19q), as part of our previous study
. The commercially purchased probes used for hybridization
were Zytolight SPEC 1p36/1q25 and 19q13/19p13 dual color
probes (Nordic BioSite, Sweden). Slides were assessed under a
fluorescence microscope (Olympus BX 50 Deutschland GmbH),
and a minimum of 100 non-overlapping nuclei was calculated for
each hybridization. A tumor was considered deleted if .50% of
the nuclei harbored two signals of the reference probe but only one
signal of the target probe.
Antibodies for detection of GABA-A channel subunits
Based on the results from the qRT-PCR and the availability of
antibodies that could be used for immunostaining of paraffin-
embedded formalin fixed tissue, we selected the following
antibodies against four different GABA-A channel subunits: One
mouse monoclonal antibody against the a1 subunit (GABRA1,
CAB22502; Chemicon, Millipore; dilution 1:50), and three
monospecific antibodies, generated through affinity purification
of polyclonal antisera, for detection of, respectively: the c1 subunit
(GABRG1, HPA035622; Atlas antibodies, Sigma-Aldrich; dilution
1:30); the r2 subunit (GABRR2, HPA016467; Atlas antibodies,
HPA002063; Atlas antibodies, Sigma-Aldrich; dilution: 1:100)
Evaluation of immunohistochemistry
The neuropathologists in this study (EA, IA) evaluated all
imunostainings and identified immunoreactive cell types as well as
the cellular distribution of the proteins. For each sample, the entire
piece of micro tissue was examined through light microscopy at
magnification 20–406. The percentage of entrapped neurons,
identified by immunoreactivity for neuronal nuclear protein, was
estimated for all samples in the TMA. Representative areas
containing the highest density of immunoreactive cells for the
GABA-A channel subunits were used for counting $200 tumor
cells per section. The fraction of immunoreactive tumor cells
labeled with GABA-A subunit antibodies was then calculated and
graded as 0–2 (0=no positive cells; 1 (+)=few (,10%) positive
cells; 2 (++)=several ($10%) positive cells).
Characterization of the TMA
Clinical characteristics and histological and molecular charac-
terization of the tumors included in the TMA are shown in
Table 3. Of the 91 tumors included in the present study, 4 tumors
were re-classified as high-grade gliomas (3 oligoastrocytomas grade
III, 1 astrocytoma grade III). The remaining tumors consisted of
astrocytoma grad II (n=42), oligodendroglioma grade II (n=27)
or oligoastrocytoma grade II (n=18). Histological analysis
revealed a fraction of 0–5% of all cells to consist of entrapped
neurons (mean percentage in all sections 2.7%). Molecular
characterization confirmed the presence of mutated IDH1
R132H protein in 76/91 tumor samples (84%) and LOH 1p/
19q in 26/91 tumor samples (29%) (Table 3).
Statistical analysis comparing mRNA levels of individual
GABA-A channel subunits between different types of gliomas
was carried out using SigmaPlot v11 (Systat Software Inc., USA),
and assessed by Kruskal-Wallis ANOVA on ranks, with signifi-
cance level set to p,0.05. The spearman correlation test was used
to calculate the coefficients between a1, c1, r2 and h subunit
protein expression and was performed in JMP statistical software,
version 5.0.1a (SAS Institute Inc., Cary, North Carolina, USA).
Survival curves were plotted according to the Kaplan-Meier
method (product-limit method) and the log-rank probability test
(Mantel-Cox) estimated the prognostic value of each specific
clinical, histological, molecular or radiological parameter in the
univariate analysis. The Cox proportional hazard model was used
to calculate the impact of each prognostic factor in the
multivariate analysis. Stepwise exclusion of variables was used to
achieve a model with as few variables as possible. The level for
confounders to be removed from the model when adjusted for
variables already in the model (P-to-remove) was set to .0.1. The
natural logarithm (ln) of the cumulative hazard plots was used to
confirm the assumption of the proportional hazard functions.
Statistical calculations were performed in JMP, version 5.0.1a
(SAS Institute Inc., Cary, North Carolina, USA).
for each of the 17 subunits, showing quantitative mRNA
levels between astrocytomas (n=6) and oligodendrogli-
omas grade II (n=6). The normalized mRNA expression of
each target gene relative to a reference gene TATA-binding
protein (TBP) was calculated using the 22DCtmethod.
Detailed presentation of the RT-PCR results
GABA-A Channel in Human Glioma
PLoS ONE | www.plosone.org9May 2012 | Volume 7 | Issue 5 | e37041
Figure S2 Download full-text
sults for each of the 17 subunits, showing quantitative
mRNA levels between astrocytomas (n=5) and oligo-
dendrogliomas grade III (n=5). The normalized mRNA
expression of each target gene relative to a reference gene
TATA-binding protein (TBP) was calculated using the 22DCt
Detailed presentation of the RT-PCR re-
units in respectively peri-tumoral tissue (n=5), gliomas
grade II (n=12), gliomas grade III (n=10), and glio-
blastomas (GBM) (n=7).
Relative mRNA levels for 17 GABARA sub-
used for qRT-PCR.
Primer sequences of GABA-A channel subunits
The authors thank IngMarie Olsson for valuable help with preparation of
the tissue microarrays.
Conceived and designed the experiments: AS ZJ EA BB. Performed the
experiments: ZJ TE HP PHE EA. Analyzed the data: AS ZJ HP TE IA AD
EA BB. Contributed reagents/materials/analysis tools: AS MN PHE FP
EA BB. Wrote the paper: AS. Edited the manuscript: ZJ HP MN BB.
Provided clinical expertise: IA EA.
1. Louis DN, Ohgaki H, Wiestler OD, Cavenee WK, Burger PC, et al. (2007) The
2007 WHO classification of tumours of the central nervous system. Acta
Neuropathol 114: 97–109.
Ohgaki H, Kleihues P (2005) Epidemiology and etiology of gliomas. Acta
Neuropathol 109: 93–108.
Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, et al. (2005)
European Organisation for Research and Treatment of Cancer Brain Tumor
and Radiotherapy Groups; National Cancer Institute of Canada Clinical Trials
Group. Radiotherapy plus concomitant and adjuvant temozolomide for
glioblastoma. N Engl J Med 352: 987–996.
Soffietti R, Baumert BG, Bello L, von Deimling A, Duffau H, et al. (2010)
Guidelines on management of low-grade gliomas: report of an EFNS-EANO
Task Force. Eur J Neurol 17: 1124–1133.
Birnir B, Korpi ER (2007) The impact of subcellular localization and
intracellular neuronal proteins on properties of GABA(A) receptors. Curr
Pharm Des 13: 3169–3177.
Young SZ, Bordey A (2009) GABA’s control of stem and cancer cell
proliferation in adult neuronal and peripheral niches. Physiology (Bethesda)
Olsen RW, Sieghart W (2008) International Union of Pharmacology. Subtypes
of c- Aminobutyric Acid A Receptors: Classification on the Basis of Subunit
Composition, Pharmacology, and Function. Pharmacol Rev 60: 243–260.
Semyanov A, Walker MC, Kullmann DM (2003) GABA uptake regulates
cortical excitability via cell type-specific tonic inhibition. Nat Neurosci 6:
Farrant M, Nusser Z (2005) Variations on an inhibitory theme: phasic and tonic
activation of GABAAreceptors. Nat Rev Neurosci 6: 215–229.
10. Bureau M, Laschet J, Bureau-Heeren M, Hennuy B, Minet A, et al. (1995)
Astroglial cells express large amounts of GABAA receptor proteins in mature
brain. J Neurochem 65: 2006–2015.
11. Colombo JA, Bentham C (2006) Immunohistochemical analysis of subcortical
white matter astroglia of infant and adult primate brains, with a note on resident
neurons. Brain Res 1100: 93–103.
12. Bjursto ¨m H, Wang J, Ericsson I, Bengtsson M, Liu Y, et al. (2008) GABA, a
natural immunomodulator of T lymphocytes. J Neuroimmunol 205: 44–50.
13. Bhat R, Axtell R, Mitra A, Miranda M, Lock C, et al. (2010) Inhibitory role for
GABA in autoimmune inflammation. Proc Natl Acad Sci U S A 107:
14. Aronica E, Redeker S, Boer K, Spliet WG, van Rijen PC, et al. (2007) Inhibitory
networks in epilepsy-associated gangliogliomas and in the perilesional epileptic
cortex. Epilepsy 74: 33–44.
15. Jussofie A, Reinhardt V, Kalff R (1994) GABA binding sites: their density, their
affinity to muscimol and their behaviour against neuroactive steroids in human
gliomas of different degrees of malignancy. J Neural Transm Gen Sect 96:
16. Labrakakis C, Patt S, Hartmann J, Kettenmann H (1998) Functional GABA(A)
receptors on human glioma cells. Eur J Neurosci 10: 231–238.
17. Synowitz M, Ahmann P, Matyash M, Kuhn SA, Hofmann B, et al. (2001)
GABA(A)-receptor expression in glioma cells is triggered by contact with
neuronal cells. Eur J Neurosci 14: 1294–1302.
18. Pignatti F, van den Bent M, Curran D, Debruyne C, Sylvester R, et al. (2002)
Prognostic factors for survival in adult patients with cerebral low-grade glioma.
J Clin Oncol 20: 2076–2084.
19. Von Deimling A, Korshunov A, Hartmann C (2011) The next generation of
glioma biomarkers: MGMT methylation, BRAF fusions and IDH1 mutations.
Brain Pathol 21: 74–87.
20. Schwirtlich M, Emri Z, Antal K, Ma ´te ` Z, Katarova Z, Szabo G (2010) GABA(A)
and GABA(B) receptors of distinct properties affect oppositely the proliferation
of mouse embryonic stem cells through synergistic elevation of intracellular
Ca(2+). FASEB J 24: 1218–1228.
21. Uusi-Oukari M, Korpi ER (2010) Regulation of GABAA Receptor Subunit
Expression by Pharmacological Agents. Pharmacol Rev 62: 97–135.
22. Capogna M, Pearce RA (2010) GABAA,slow: causes and consequences. Trends in
Neuroscience 34: 101–112.
23. Belelli D, Harrison NL, Maguire J, Macdonald RL, Walker MC, et al. (2009)
Extrasynaptic GABAA receptors: form, pharmacology, and function. J Neurosci
24. Rosas-Arellano A, Parodi J, Machuca-Parra AI, Sa ´nchez-Gutie ´rrez A,
Inestrosa NC, et al. (2011) The GABA(A)r receptors in hippocampal
spontaneous activity and their distribution in hippocampus, amygdala and
visual cortex. Neurosci Lett 500: 20–25.
25. Feigenspan A, Wa ¨ssle H, Bormann J (1993) Pharmacology of GABA receptor
Cl- channels in rat retinal bipolar cells. Nature 361: 159–162.
26. Pan Y, Ripps H, Qian H (2006) Random assembly of GABA rho1 and rho2
subunits in the formation of heteromeric GABA(C) receptors. Cell Mol
Neurobiol 26: 289–305.
27. Martı ´nez-Delgado G, Estrada-Mondrago ´n A, Miledi R, Martı ´nez-Torres A
(2010) An update on GABAr Receptors. Curr Neuropharmacol 8: 422–433.
28. Bormann J, Feigenspan A (1995) GABAC receptors. Trends Neurosci 18:
29. Enz R, Cutting GR (1999) GABAC receptor rho subunits are heterogeneously
expressed in the human CNS and form homo- and heterooligomers with distinct
physical properties. Eur J Neurosci 11: 41–50.
30. Shaw EG, Daumas-Duport C, Scheithauer BW, Gilbertson DT, O’Fallon JR, et
al. (1989) Radiation therapy in the management of low-grade supratentorial
astrocytomas. J Neurosurg 70: 853–861.
31. Okamoto Y, Di Patre PL, Burkhard C, Horstmann S, Jourde B, et al. (2004)
Population-based study on incidence, survival rates, and genetic alterations of
low-grade diffuse astrocytomas and oligodendrogliomas. Acta Neuropathol 108:
32. Yang HJ, Kim JE, Paek SH, Chi JG, Jung HW, Kim DG (2003) The
significance of gemistocytes in astrocytoma. Acta Neurochir (Wien) 145:
33. Angulo MC, Le Meur K, Kozlov AS, Charpak S, Audinat E (2008) GABA, a
forgotten gliotransmitter. Prog Neurobiol 86: 297–303.
34. Schmittgen TD, Livak KJ (2008) Analyzing real-time PCR data by the
comparative C(T) method. Nat Protoc 3: 1101–1108.
35. Qu M, Aronica E, Boer K, Fa ¨llmar D, Kumlien E, et al. (2009) DLG3/SAP102
protein expression in malformations of cortical development: a study of human
epileptic cortex by tissue microarray. Epilepsy Res 84: 33–41.
36. Elsir T, Qu M, Berntsson SG, Orrego A, Olofsson T, et al. (2011) PROX1 is a
predictor of survival for gliomas WHO grade II. Br J Cancer104: 1747–1754.
37. Uhle ´n M, Ponte ´n F (2005) Antibody-based proteomics for human tissue
profiling. Mol Cell Proteomics 4: 384–393.
38. Capper D, Zentgraf H, Balss J, Hartmann C, von Deimling A (2009)
Monoclonal antibody specific for IDH1 R132H mutation. Acta Neuropathol
39. Paavilainen L, Werne ´rus H, Nilsson P, Uhle ´n M, Hober S, et al. (2008)
Evaluation of monospecific antibodies: a comparison study with commercial
analogs using immunohistochemistry on tissue microarrays. Immunohistochem
Mol Morphol 16: 493–502.
GABA-A Channel in Human Glioma
PLoS ONE | www.plosone.org10May 2012 | Volume 7 | Issue 5 | e37041